High frequency of multidrug-resistant Mycobacterium tuberculosis isolates in Georgetown, Guyana
Nikolai Menner, Jutta Wagner, Helmut Hahn, and Ralf Ignatius (corresponding author), Department of Medical Microbiology and Immunology of Infection, Charité– University Medicine Berlin, Benjamin Franklin Campus, Hindenburgdamm 27, 12203 Berlin, Germany. Tel.: +49 30 8445 3620; Fax: +49 30 8445 3830; E-mail: firstname.lastname@example.org
Ingo Günther, Faculty of Health Sciences, School of Medicine, University of Guyana, Turkeyen Campus, East Coast Demerara, P.O. Box 101110, Georgetown, Guyana.
Helmut Orawa, Department of Biostatistics and Clinical Epidemiology, Charité– University Medicine Berlin, Benjamin Franklin Campus, Hindenburgdamm 30, 12200 Berlin, Germany.
Andreas Roth, Institute for Microbiology, Immunology, and Laboratory Medicine, Hospital Zehlendorf-Heckeshorn, Zum Heckeshorn 33, 14109 Berlin, Germany
Indal Rambajan and Shamdeo Persaud, Ministry of Health, Guyana, 11 Brickdam Stabroek, Georgetown, Guyana.
Emergence of multidrug-resistant (MDR) Mycobacterium tuberculosis isolates constitutes a threat to public health worldwide. This study aimed at acquiring first epidemiological data for Guyana. Thirty-six M. tuberculosis isolates from patients of the Georgetown Chest Clinic were subjected to susceptibility testing on solid agar and in broth media. Resistance to at least one first-line drug was observed in 8 (22.2%, 95% confidence interval 8.3–36.1%) and simultaneous resistance to rifampicin and isoniazid (MDR) in 4 (11.1%, 95% confidence interval 0.6–21.6%) of the 36 isolates. The risk of infection with resistant isolates was significantly related to earlier antituberculosis therapy (P = 0.040). These data indicate a high proportion of resistant M. tuberculosis isolates in Guyana and call for the implementation of control strategies based on an improved laboratory diagnosis of TB.
Despite recent advances in the control of tuberculosis (TB), this disease still remains endemic in Latin America (reviewed in Pelly et al. 2004). In conjunction with the AIDS pandemic, more people are at risk of developing mycobacterial infections, which in Latin American countries are caused mostly by Mycobacterium tuberculosis, whilst only a minority of patients acquire non-tuberculous mycobacteria (NTM). As TB often affects young people, particularly in case of HIV co-infection, this disease also has dramatic social and economic consequences. The introduction of DOTS (directly observed treatment, short course) has contributed in many countries to a better control of TB. Control campaigns, however, must be based on the knowledge of the prevailing regional resistance of M. tuberculosis to standard antituberculosis chemotherapeutic drugs, which enables physicians to expand the initial regimen of drugs or choose alternative drugs.
As the emergence of multidrug-resistant (MDR) isolates of M. tuberculosis appears to be a worldwide phenomenon (Espinal et al. 2001), several studies have recently been performed in various countries to evaluate the susceptibility of regional isolates of M. tuberculosis. The lack of relevant data on this topic for Guyana is mainly due to the fact that in Guyana mycobacteria are not routinely isolated from patients’ samples and subjected to susceptibility testing. Nevertheless, TB remains a major health problem in this country with a population of approximately 765 000, as in 2003 the estimated incidence rates and death rate for all cases of TB including HIV-positive cases were 130/100 000 (57 smear-positives/100 000) and 24/100 000, respectively. With a total prevalence of 184 cases per 100 000 subjects, these numbers are around threefold higher than those reported for 1990 (WHO 2005). In a first effort to elucidate the current MDR-TB situation in Guyana, we collected M. tuberculosis isolates in Guyana from sputum samples from suspected patients. Susceptibility of these isolates was tested against the five first-line antituberculosis drugs as well as against three second-line drugs that could potentially be alternatives in the treatment of TB, i.e., prothionamide, capreomycin and cycloserine.
Materials and methods
After the implementation of a small mycobacteria laboratory, expectorated sputum samples were consecutively collected from 166 patients with suspected TB admitted to the Georgetown Chest Clinic, Guyana, between February 1st and April 30th, 2001. This clinic is the central institution for the diagnosis of TB with the responsibilities of distributing free treatment to the patients and documentation of TB cases. All patients were interviewed as to previous or present antituberculosis treatment, and possible risk factors for the development of drug-resistance, e.g., HIV-infection (Campos et al. 2003). Procedures of growth, differentiation and susceptibility testing of mycobacteria followed standard protocols. Briefly, samples were decontaminated with NaOH and N-acetyl-L-cysteine and cultured on Stonebrink and Loewenstein–Jensen agar (BAG, Lich, Germany) for 12 weeks, whereas cultures in liquid media could not be performed. Smears were made from all samples and stained (Ziehl-Neelsen method) to detect acid-fast bacilli (AFB). Growth of AFB was verified by Ziehl-Neelsen staining and bacterial isolates were first differentiated using the AccuProbe rRNA hybridization system (Gen-Probe, bioMérieux, Nürtingen, Germany), which reliably identifies mycobacteria of the M. tuberculosis complex (Evans et al. 1992). All isolates of the M. tuberculosis complex were further specified by routine biochemical reactions (niacin, nitrate reduction and pyromucic acid hydrazide) and genetically in a commercially available DNA strip assay that reliably distinguishes mycobacteria of the MTB complex (Genotype MTBC; Hain Lifescience GmbH, Nehren, Germany) (Richter et al. 2003). NTM were genetically differentiated as previously described (Roth et al. 2000). Susceptibility of the M. tuberculosis isolates against rifampicin (MIC ≥32μg/ml), isoniazid (MIC ≥0.25 μg/ml), ethambutol (MIC ≥1 μg/ml), and streptomycin (MIC ≥4 μg/ml) was determined following the standard agar dilution technique on Loewenstein–Jensen agar (Heipha Diagnostika, Eppelheim, Germany). Data were confirmed in the BBL MGIT AST SIRE system (isoniazid, MIC ≥0.1 μg/ml; rifampicin, MIC ≥1.0 μg/ml; ethambutol, MIC ≥3.5 μg/ml; streptomycin, MIC ≥0.8 μg/ml) (Becton Dickinson, Heidelberg, Germany) (Rusch-Gerdes et al. 1999). Susceptibility to prothionamide, capreomycin and cycloserin (all MIC ≥16 μg/ml) was tested by agar dilution only, while susceptibility to pyrazinamide was tested in a standard pyrazinamidase assay (Wayne 1974). Three control isolates (M. tuberculosis H37Rv and two M. tuberculosis isolates with defined resistances) were constantly tested in parallel. Data were statistically analysed using the chi-square and exact Fisher tests.
AFB were detected microscopically in smears from 49 of 203 sputum samples (24.1%) obtained from 166 patients. Fifty-six cultures from 47 patients yielded growth of mycobacteria. Isolates from 36 patients were biochemically and genetically identified as M. tuberculosis, while NTM were isolated from 12 patients. One sputum sample contained both M. tuberculosis and NTM. Further differentiation of NTM revealed Mycobacterium gordonae, Mycobacterium intracellulare, Mycobacterium abscessus, Mycobacterium fortuitum (two times), Mycobacterium peregrinum, Mycobacterium kansasii, and five isolates consisted of more than one species of NTM. Notably, 45 of the 49 smear-positive samples (91.8%) yielded growth of M. tuberculosis.
Susceptibility testing was performed for all 36 M. tuberculosis isolates (Table 1). Eight of 36 isolates (22.2%, 95% confidence interval 8.3–36.1%) displayed either single or multiple resistance to the first-line substances tested. Resistance to isoniazid or rifampicin was most frequently detected: in 6 (16.7%) and 5 (13.9%) of the 36 isolates, respectively. Three isolates each (8.3%) were resistant to ethambutol or pyrazinamide. Based on their simultaneous resistance to isoniazid and rifampicin, four of the eight isolates (11.1% of all 36 M. tuberculosis isolates, 95% confidence interval 0.6–21.6%) were judged as MDR. At the time of the study, TB patients in Guyana were routinely treated with the combination of isoniazid, rifampicin, ethambutol and pyrazinamide, while streptomycin was only given when treatment failure was suspected. Consequently, we detected only one streptomycin-resistant isolate (2.8%). Resistance data to first-line drugs obtained using the MGIT system and those by agar dilution technique were identical. None of the 36 isolates exhibited resistance to prothionamide, whereas 2 (5.6%) and 3 (8.3%) isolates were resistant to capreomycin and cycloserine, respectively, suggesting these drugs as potential alternatives for the treatment of TB in Guyana.
Table 1. Resistance patterns of M. tuberculosis isolates (n = 36)
| 2||R||R||R||S||S||S||S||S||+, +|
| 1||R||S||S||R||R||S||S||S|| |
| 1||R||S||S||S||S||S||S||S|| |
| 1||S||R||S||S||S||S||S||S|| |
| 1||S||S||S||S||R||S||S||S|| |
| 3||S||S||S||S||S||S||S||R|| |
| 1||S||S||S||S||S||S||R||S|| |
A risk factor for the presence of resistance was previous antituberculosis treatment, as resistant isolates were significantly more often found in previously treated (5 of 11 patients, 45.5%) than in treatment-naive patients (3 of 25 patients, 12%) (P = 0.040). Similarly, prior treatment was a risk factor for MDR when tested versus any resistance (but not MDR) versus no resistance (P = 0.044). Interestingly, there was also a significant association between age (four groups) and resistance (yes/no) (P = 0.040). The proportions of resistant isolates amongst all isolates of the respective age group were 0/11 (20–29 years), 6/13 (30–39 years), 2/9 (40–49 years) and 0/3 (50–59 years). Due to the lack of reliable results of HIV-testing for the patients included in the study (the HIV-status could only be judged based on the information obtained during the initial interview) no solid data could be obtained regarding the possible correlation of MDR-TB and HIV-infection.
Even though the number of isolates that could be collected for this study was relatively small, the data indicate the existence of a considerable proportion of MDR M. tuberculosis isolates in Guyana. If a comparable range is subsequently confirmed through more extensive studies, it would be considerably higher than that reported for many other countries (Espinal et al. 2001). Nevertheless, in a recent study on 170 M. tuberculosis isolates from patients from the state of Rio Grande do Sul, Brazil (Almeida da Silva et al. 2001), resistance to at least one TB drug was detected in 14% and 73% of treatment-naive and pre-treated patients, respectively, as compared to 12% and 45% in our study. In Ecuador, MDR was observed in 8.7% of isolates obtained from 161 patients who had had no prior anti-TB treatment compared to 16.7% of 60 isolates from patients with previous treatment (Mertz et al. 2000).
Although DOTS had been implemented in Guyana by 2001, many patients still received their weekly medication, in cases of longer distances to the clinic even the medication for 1 month, without the possibility to control the treatment outcome. Accordingly, in many cases the use of the drugs could not be monitored, and we found that the risk of resistant TB was significantly correlated with a history of previous antituberculosis therapy. The importance of effective programs for the control of MDR-TB in resource-limited settings, however, has recently been emphasised (Mukherjee et al. 2004), and studies in Peru demonstrated the success of community-based programs for the treatment of TB and MDR-TB (Mitnick et al. 2003). In contrast, poor management of control programs can favour the spread of MDR-TB isolates (Laserson et al. 1998). Interestingly, a new assay for susceptibility testing of M. tuberculosis suitable for the use in resource-limited settings has recently been described by the Peruvian TB working group (Moore et al. 2004), but monitoring of treatment failure can also be helpful for identification of patients with MDR-TB (Becerra et al. 2000). Unfortunately, monitoring of the outcome of DOTS could not be implemented in Guyana yet (WHO 2005). Notably, a recent study on drug resistance profiles of 1680 M. tuberculosis isolates in Lima, Peru (Timperi et al. 2005), demonstrated a MDR-TB prevalence of 55% in the population studied and significantly increased numbers of isolates resistant to at least five first-line drugs (from 29% to 37%) for the years 1996–2001 despite a well-established model DOTS program since 1991. Whereas primarily testing for resistance to INH and RMP may be useful for the initial selection of a defined treatment regimen in resource-limited settings, a local laboratory capacity for validated drug-sensitivity testing appears to be essential for the development of individualized as well as standardized treatment as well as re-treatment regimens.
In conclusion, the data of the present study indicate high frequency of MDR-TB in Guyana and urgently call for (i) the establishment of a reference laboratory with the necessary equipment and personnel for TB diagnosis and susceptibility testing in Guyana, (ii) a more comprehensive study on the occurrence of drug-resistant M. tuberculosis isolates in Guyana and (iii) the improvement of present TB control strategies, e.g., the monitoring of treatment outcome.
We thank Hermann Feldmeier for helpful discussions and critical reading of the manuscript, as well as Heike Mellenthin and Dagmar Piske for excellent technical assistance. N.M. gratefully acknowledges a travel grant from the Verein der Freunde und Förderer des UKBF e.V.